R is given as a function of the Reynolds' number vD/v from the data of Wieselsberger 

 and Relf. v is the speed, D the diameter of the rod, and V the kinematic viscosity. 



From Fig. 2 it is seen that the experimental spots approach the computed 

 curve with increasing rod length. The hump determined by the spots for the four 

 foot rod over the range 1.5 to 2.5 ft. /sec. was correlated with a maximum lateral 

 vibration of the rod. A less pronounced hump for the three foot rod over the range 



2.2 to 3.2 ft. per sec. is of the same origin. These facts imply that the stream 

 impresses upon the rod an oscillating force whose frequency increases with the speed. 

 For the shorter rod, which had a higher natural frequency, reached a condition of 

 maximum vibration (resonance) at a higher speed. For further verification a four 



foot brass rod (Fig. 3) was tested. Since the elastic modulus for brass is 



7 7 



1.3 x 10 , and that for steel is 3.0 x 10 , while the masses are in the ratio of 



9 to 8, the ratio of the resonance frequency of brass to that of steel is 



*■£ x -*- = 0.62. Consequently the hump should occur at a lower speed for brass, as 



is observed to be the case. Some residual vibration was observed to the highest 

 speeds. To this may be attributed the small but uniform deviation of the spots 

 from the computed curve in Fig. 3. 



Conclusion: 



In agreement with the results of Relf and Powell, the function R sin a $ 

 furnishes a good approximation to the law of force normal to an infinite rod at an 

 angle $ with a stream. 



